JP2007051582A - Fuel-injection control system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a fuel-injection control system by which correct commands can be acquired before and after learning a correction. <P>SOLUTION: A command storage 3 stores a pre-learning command for the time before a correction-learning means 5 learns the correction, and a post-learning command for the time after the correction-learning means 5 has learned the correction. A command injection-volume-determining means 4 refers to the pre-learning command at the time before the learning, and it refers to the post-learning command at the time after the learning. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a multi-injection fuel injection control system, and more particularly to a fuel injection control system that can obtain an appropriate command before and after learning a correction amount.

  When a fuel injection unit is provided for each cylinder of a diesel engine and the injection amount is controlled by controlling the energization time to the fuel injection unit, the fuel injection control system performs injection based on the current engine control parameter. The command injection amount and the number of injections to be performed are determined. At this time, in order to reduce the complexity of calculation, it is common to store the command injection amount to be injected by the fuel injection means in accordance with engine control parameters in a command storage means called a map. The map stores a command injection amount to be injected by the fuel injection means in correspondence with the engine control parameter. By referring to the map based on the engine control parameter, the command injection amount can be obtained.

  The fuel injection control system does not inject the entire command injection amount to be injected only once in one combustion cycle, but appropriately performs preliminary injection called pilot injection or after injection before and after the main injection. This is called multi-injection. In the multi-injection, the command injection amount is injected in a plurality of times. Command storage means that maps the injection pattern (command injection amount and number of injections) is called a multi-injection pattern map.

  From the basic principle that the injection amount is proportional to the energization time, the command injection amount is given by the energization time. However, in actuality, there is a variation (individual difference) in the ratio between the injection amount and the energization time depending on the individual fuel injection means, so that the actual injection amount actually injected by the fuel injection means is the same as the command injection amount. Every correction is required.

  Although the required correction amount varies depending on the individual fuel injection means, the required correction amount does not vary greatly within a short time in the same individual fuel injection means. Therefore, conventionally, the correction amount for the command injection amount is learned so that the actual injection amount actually injected by the fuel injection means is equal to the command injection amount, and after the learning, the command injection amount is corrected by the learned correction amount. It has become. Since the command injection amount is given by the energization time, the correction amount is also given by the energization correction time for shortening or extending the energization time of the command injection amount.

  The learned correction amount is stored in a nonvolatile memory. Thereby, the learned correction amount is retained even after the power is turned off, and the stored correction amount can be used without learning again when the power is turned on next time.

  In addition, the fuel injection control system uses an engine speed (hereinafter referred to as a target engine speed) that is the target of the engine speed (hereinafter referred to as an actual engine speed) at which the engine is actually rotating during idle operation. In order to match the engine speed), a feedback amount is determined by multiplying the deviation between the target engine speed and the actual engine speed by a proportional coefficient (hereinafter referred to as an idle feedback coefficient), and the feedback amount is used as the command injection quantity. By superimposing, the target engine speed is corrected so that the actual engine speed approaches the target engine speed.

JP 2004-11511 A Japanese Patent Laid-Open No. 2000-8908 JP 2005-16486 A

  In the conventional fuel injection control system, the idle feedback coefficient and the multi-injection pattern map are the same before and after learning the correction amount. However, the idle feedback coefficient and the multi-injection pattern map are naturally created on the assumption that the injection is performed accurately (the actual injection amount is the same as the command injection amount). That is, it matches the state after learning. Therefore, the fact that the idle feedback coefficient and the multi-injection pattern map are the same before and after learning the correction amount means that the injection is not accurately performed (the actual injection amount is not in accordance with the command injection amount). This means that the idle feedback coefficient and the multi-injection pattern map are used before learning.

  Conventionally, this has caused idling hunting and fluctuation between cylinders. When idling hunting or cylinder-to-cylinder fluctuations occur, learning becomes impossible.

  Conversely, even if an idle feedback coefficient or a multi-injection pattern map that matches the state in which the injection before learning is not accurately performed is created, the injection is accurately performed this time. If it is used after learning, the learned meaning is halved.

  Further, if an extremely large feedback amount is given during idling, the engine may stop or oscillate (the engine speed may fluctuate without being stabilized). On the other hand, if the feedback amount is too small, it takes time for the engine speed to stabilize at the target value.

  In addition, it has been described that the necessary correction amount does not vary greatly within a short time in the same fuel injection means, but the state of the fuel injection means may change over a long span such as the life of the fuel injection means. Conceivable. Therefore, if the learned correction amount is stored in the nonvolatile memory and used continuously, the state of the fuel injection means and the correction amount may not match.

  Accordingly, an object of the present invention is to provide a fuel injection control system capable of solving the above-described problems and obtaining an appropriate command both before and after learning a correction amount.

  In order to achieve the above object, the present invention provides a command storage means for storing the command injection amount and the number of injections to be injected by the fuel injection means corresponding to the engine control parameter, and the command storage based on the current engine control parameter. The command injection amount determining means for determining the command injection amount and the number of injections with reference to the means, and the correction amount for the command injection amount are learned so that the actual injection amount actually injected by the fuel injection means is equal to the command injection amount. In the fuel injection control system provided with the correction amount learning means, the command storage means includes a pre-learning command for the pre-learning time before the correction amount learning means learns the correction amount, and the correction amount learning means. The post-learning command for after learning after learning the correction amount is stored, and the command injection amount determining means refers to the pre-learning command before learning, and is used for after learning after learning. It is intended to refer to the decree.

  The average number of injections in the pre-learning command for the same engine control parameter may be smaller than the average number of injections in the post-learning command.

  During idling, the target engine speed is compared with the actual engine speed, the feedback amount is determined by multiplying the deviation by a proportional coefficient, and the feedback amount is superimposed on the command injection amount. Idle feedback means so that the number approaches the target engine speed, and the idle feedback means learns larger than the pre-learning proportionality coefficient for before learning and the pre-learning proportionality coefficient for after learning. It is also possible to store the post-proportional coefficient, determine the feedback amount using the pre-learning proportional coefficient before learning, and use the post-learning proportional coefficient after learning.

  When the learning is completed, the correction amount learning unit stores in the memory whether or not learning has been performed together with the learned correction amount, and performs learning when it is determined that learning has been performed based on the data indicating whether or not learning has been performed. Instead, the command injection amount determination means determines whether it is before learning or after learning based on the data indicating whether learning has been completed, and whether or not a re-learning condition set in advance is satisfied. Re-learning determination means for determining whether or not when the re-learning determination means has become a re-learning condition, the data indicating whether or not the learning has been completed is set to a state before learning. A pre-learning means for returning the correction amount learning means and the command injection amount determining means before the learning may be provided.

  The present invention exhibits the following excellent effects.

  (1) An appropriate command can be obtained both before and after learning the correction amount.

  Hereinafter, an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

  As shown in FIG. 1, the fuel injection control system 1 according to the present invention is a command storage means for storing the command injection amount and the number of injections that the fuel injection means 2 should inject during the combustion cycle in accordance with engine control parameters. 3, command injection amount determination means 4 for determining the command injection amount and the number of injections with reference to the command storage means 3 based on the current engine control parameters, and the actual injection amount actually injected by the fuel injection means 2 In the fuel injection control system 1 provided with the correction amount learning means 5 for learning the correction amount for the command injection amount so that the value becomes the command injection amount, the command storage means 3 stores the correction amount by the correction amount learning means 5. A pre-learning command for before learning, which is before learning, and a post-learning command for after learning, after the correction amount learning means 5 has learned the correction amount. It means 4 are those at the time before learning with reference to the command for pre-learning, and sometimes to refer to a post-training instruction after learning.

  The engine control parameter may be used by combining any number of known quantities conventionally input to the fuel injection control system 1 such as the engine speed, engine torque, accelerator opening, air fuel ratio, exhaust gas recirculation amount, and the like. Here, for simplicity, only the engine speed and torque will be described as an example.

  As the fuel injection means 2, various fuel injection means whose injection amount is substantially proportional to the energization time may be used. Here, a solenoid-type injector that opens and closes the valve of the injection nozzle with a solenoid is used.

  The correction amount learning means 5 is a well-known one, and for the injection cylinder indicated by the cam angle sensor (not shown), the command injection amount determined by the command injection amount determination means 4 to the fuel injection means 2 and the engine speed. A correction amount can be obtained by comparing the engine speed of the injection cylinder indicated by a sensor (not shown) and estimating the excess or deficiency of the actual injection amount in the fuel injection means 2.

  The command storage unit 3, the command injection amount determination unit 4, the correction amount learning unit 5, the idle feedback unit 7 and the learning pre-return unit 9, which will be described later, are realized by software in an engine control computer called an ECU 6.

  The command storage means 3 is divided into a pre-learning map 3a for storing pre-learning commands and a post-learning map 3b for storing post-learning commands. Each map is a multi-injection pattern map that stores a command injection amount, the number of injections, an injection amount pattern, an injection timing, and an injection interval. Here, for simplicity, only the number of injections will be described as an example.

  The fuel injection control system 1 compares a target engine speed with a target engine speed at the time of idling, determines a feedback amount by multiplying the deviation by a proportional coefficient (idle feedback coefficient), and the feedback amount. Is superimposed on the command injection amount so that the actual engine speed approaches the target engine speed, and the idle feedback means 7 has a relatively small pre-learning time before learning. The proportional coefficient and the comparatively large post-learning proportional coefficient for after learning are stored, the pre-learning proportional coefficient is used before learning, and the post-learning proportional coefficient is used after learning to determine the feedback amount. It is like that.

  In the fuel injection control system 1, the correction amount learning means 5 stores the learned correction amount in the nonvolatile memory 8 when learning is completed and also indicates data that has been learned (for example, a learned flag). Is stored in the non-volatile memory 8 and learning is not performed while the learned flag is present, and the command injection amount determining means 4 and the idle feedback means 7 are either before learning or after learning depending on the learned flag. It is as a judgment. Note that the method of storing whether or not learning has been performed as data is not limited to the learned flag. When the learning value is represented by a 2-digit hexadecimal number, when the learning value is represented as “FF” in an unlearned state, the learning value is read and the value is “FF”. Sometimes it can be determined that it has not been learned. Since this method does not require a learned flag, it can save memory capacity.

  The fuel injection control system 1 has re-learning determination means (not shown) for determining whether or not a re-learning condition set in advance is satisfied, and the re-learning determination means becomes the re-learning condition. When it is determined that the learning has been completed, the learning amount return means 9 for returning the correction amount learning means 5 and the command injection amount determination means 4 to the pre-learning state by setting the data indicating whether or not learning has been performed to a pre-learning state. Is provided.

  The nonvolatile memory 8 can be configured by an EEPROM, a flash memory, or the like.

  FIG. 2A shows a pre-learning map, and FIG. 2B shows a post-learning map. As shown in the figure, in any map, since the engine control parameters are only the engine speed and the engine torque, each map can be expressed two-dimensionally. By referring to the column located in the column of the desired engine speed (rpm) in the desired engine torque (N) row, the number of injections (times) stored in that column can be read. If the engine torque and engine speed are intermediate values in the rows and columns of these maps, round the engine torque and engine speed to the row and column values and refer to the map, or check the engine torque and engine speed. The number of injections may be obtained by an approximate expression from the values on both sides sandwiching.

  Comparing the same field of two maps, the number of injections in the pre-learning map is less than or equal to the number of injections in the post-learning map. That is, the number of injections in the pre-learning command for the same engine control parameter is smaller than the number of injections in the post-learning command. That is, the average value of the injection amount in the map, more specifically, the average value of the number of injections of the map during normal operation in which learning is performed is smaller than the average value of the number of injections of the map after learning. The reason will be described with an example.

Now, an engine with a displacement of 1700 cm ³ is assumed. The injection amount required for one combustion cycle of one cylinder at the time of idling is 4 mm ³ . Before learning, injection is not performed accurately (the actual injection amount is not the same as the command injection amount). Suppose that there is an individual whose fuel injection means 2 has an actual injection amount that is 1 mm ³ greater than the command injection amount per injection. When this fuel injection means 2 is injected three times, the total increases by 3 mm ³ . That is, the actual injection amount is 3 mm ³ more than the command injection amount in one combustion cycle. This increased amount is considerably larger than the command injection amount 4 mm ³ , so that the idle operation cannot be maintained. On the contrary, in an individual in which the injection amount per one time is small, the actual injection amount of one combustion cycle is very small, and it may be impossible to maintain the idling operation.

On the other hand, in the present invention, the number of injections before learning is made smaller. For example, if only one injection is used in the previous example, the actual injection amount is only 1 mm ³ greater than the command injection amount in one combustion cycle. Therefore, idle operation can be maintained. Since the idling operation can be maintained, the correction amount can be learned at the idling time.

  However, after learning, the injection is accurately performed (the actual injection amount is the same as the command injection amount), so that the error does not increase no matter how many injections are performed in one combustion cycle. Therefore, it is possible to increase the number of injections even during idling to perform multi-injection, which is the original purpose.

  FIG. 3 shows an equivalent circuit of the feedback amount determination process performed by the idle feedback means 7. This equivalent circuit includes a comparator 31 that compares the target engine speed and the actual engine speed, a pre-learning coefficient unit 32 that stores a pre-learning proportional coefficient, and a post-learning coefficient unit that stores a post-learning proportional coefficient. 33 and a switch 34 that switches the read values of these two memories before and after learning and outputs them as a feedback amount.

  The proportional coefficient for before learning is relatively small, whereas the proportional coefficient for after learning is relatively large. One reason is that prior to learning, the engine (during idle operation) is not stopped, that is, the engine stability is emphasized. If the proportionality factor is large, the feedback amount becomes very large when the deviation between the target engine speed and the actual engine speed is large, which may cause the engine to stop, but this is avoided by the small proportionality factor. it can.

  Another reason is that, after learning, it is important that the engine speed converges quickly to the final target idle speed and the injection quantity during idling becomes the optimum injection quantity. If the proportionality coefficient is small, the feedback amount does not become so large with respect to the deviation between the target engine speed and the actual engine speed, so the convergence is slow. However, if the proportionality coefficient is large, the convergence can be accelerated.

  FIG. 4 shows the flow of control. The operation of the fuel injection control system will be described below according to this flow.

  This control flow starts when the power is turned on (ignition key is on). Immediately after the start, in step S1, the ECU 6 reads the learned flag stored in the nonvolatile memory 8 and stores it in the work area in the ECU 6. Note that the learned flag = 0 (clear) when the ECU 6 is shipped.

  In step S2, the ECU 6 determines whether or not the learned flag is 1 (whether learning has been completed). If NO, the current time is before learning. If YES, the current time is after learning.

  If it is before learning, in step S3, the command injection amount determination means 4 determines the command injection amount and the number of injections with reference to the pre-learning map. At this time, if it is during idling, the idle feedback means 7 determines the feedback amount using the pre-learning proportionality coefficient and calculates the target engine speed. In other words, the command injection amount is obtained by using an idle feedback coefficient or a multi-injection pattern map that is optimized in a state where the injection before learning is not accurately performed (the actual injection amount is not in accordance with the command injection amount). The amount of correction for is obtained.

  In step S4, the ECU 6 determines whether or not learning by the correction amount learning means 5 has been completed. If NO, the process jumps to step S6.

  Here, the correction amount learning means 5 sets a specific engine speed for the ideal fuel injection quantity, and finishes learning with the command fuel injection quantity at which the engine speed is stabilized at the specific engine speed as the ideal fuel injection quantity. Good.

For example, in the case of an engine that rotates 700 mm at a theoretical value of 5 mm ³ , if it is an injector that injects 3 mm ³ more, if the command injection amount is gradually decreased from 5 mm ³ during learning, the command injection amount Is stable at 700 rpm at 2 mm ³ . Since the rotation is stabilized at 700 rpm, the difference between the actual injection amount and the commanded injection amount can be filled by correcting the commanded injection amount (energization time, etc.) at this time as 5 mm ³ thereafter. . It is determined that the learning is finished when the engine rotation is stabilized.

  If step S4 is YES, the learned flag is set to 1 (learned) in step S7. Proceed to step S6.

  In step S6, the ECU 6 determines whether or not the ignition key is off. Since the memory cannot be backed up when the power is turned off, the memory is backed up when the ignition key is turned off. If NO, since the ignition key is on, it can be seen that the engine is running, and the process returns to step S2. If YES, the engine operation has been stopped because the ignition key is off. Therefore, in step S8, the ECU 6 stores the learned flag in the work area in the ECU 6 in the nonvolatile memory 8 to prepare for power off.

  If it is determined in step S2 that it is after learning, in step S9, the command injection amount determination means 4 refers to the post-learning map and determines the command injection amount and the number of injections. At this time, if it is during idling, the idle feedback means 7 determines the feedback amount using the post-learning proportionality coefficient and calculates the target engine speed. In other words, the command injection amount is obtained by using an idle feedback coefficient or a multi-injection pattern map that is optimized in a state in which the learned injection is accurately performed (a state in which the actual injection amount is the same as the command injection amount). The amount of correction for is obtained.

  In step S10, the deterioration determination of the fuel injection means 2 by the learning pre-return means 9 is performed. Specifically, when the deterioration determination variable exceeds a predetermined determination reference value, it is determined that the fuel injection unit 2 has deteriorated (the state has changed to the extent that relearning is necessary). Examples of the deterioration determination variable include an idle feedback integral term and a vehicle travel distance. Further, it is also possible to detect variations in idle rotation speed, detect variations in correction values between cylinders, etc., and use the magnitude of these variations as a deterioration determination variable. In addition, when the fuel injection means 2 is replaced or when a part (pump, ECM, etc.) that affects the fuel injection system is replaced, re-learning is necessary. On the other hand, information that conveys the event can be input, and this information may be determined in step S10.

  If the determination in step S10 is NO, there is no need to relearn the correction amount, and the ECU 6 jumps to step S6. If the determination in step S10 is YES, the correction amount needs to be re-learned, so after setting the learned flag = 0 (clear) in step S11, the process jumps to step S6. The effect of clearing the learned flag appears in step S2, and relearning occurs.

  As described above, in the state where the injection before learning the correction amount is not accurately performed, the command injection amount is obtained using the idle feedback coefficient or the multi-injection pattern map optimized in that state, or the command injection In the state where the correction amount for the amount is obtained and the injection after learning the correction amount is accurately performed, the command injection amount using the idle feedback coefficient and the multi-injection pattern map optimized in that state Or a correction amount for the command injection amount. That is, an appropriate command can be obtained from the command injection amount determining means 4 and the idle feedback means 7 before or after learning.

  Since the number of injections before learning is set smaller than the number of injections after learning, an error in the actual injection amount with respect to the command injection amount does not increase. In particular, idling can be maintained at idling when the command injection amount is originally small.

  Since the idle feedback means 7 uses a relatively small pre-learning proportional coefficient before learning and uses a relatively large post-learning proportional coefficient after learning, the engine rotation stability is maintained before learning and the engine rotation after learning. Rapid convergence to an appropriate number can be expected.

  Since the pre-learning return means 9 determines the deterioration of the fuel injection means 2 and clears the learned flag to return the various controls to the time before learning, the correction amount corresponding to the deterioration of the fuel injection means 2 can be re-learned.

It is a block block diagram of the fuel-injection control system which shows one Embodiment of this invention. The example of the instruction | command memory | storage means used for this invention is shown, (a) is a map before learning, (b) is a figure of the map after learning. It is an equivalent circuit diagram of the feedback amount determination process performed by the idle feedback means used in the present invention. It is a flowchart which shows the flow of control in the fuel-injection control system of this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Fuel injection control system 2 Fuel injection means 3 Command storage means 4 Command injection amount determination means 5 Correction amount learning means 6 ECU (computer)
7 Idle feedback means 8 Non-volatile memory 9 Learning return means

Claims (4)

  Command storage means for storing the command injection amount and the number of injections to be injected by the fuel injection means corresponding to the engine control parameter, and the command injection amount and the injection frequency with reference to the command storage means based on the current engine control parameter Fuel injection control comprising: command injection amount determining means for determining the correction amount; and correction amount learning means for learning a correction amount for the command injection amount so that the actual injection amount actually injected by the fuel injection means is equal to the command injection amount In the system, the command storage means includes a pre-learning instruction for before the learning that is before the correction amount learning means learns the correction amount and a post-learning time after the correction amount learning means has learned the correction amount. A post-learning command for the fuel, wherein the command injection amount determining means refers to the pre-learning command before learning, and refers to the post-learning command after learning Cum control system.   2. The fuel injection control system according to claim 1, wherein the average number of injections in the pre-learning command for the same engine control parameter is smaller than the average number of injections in the post-learning command.   During idling, the target engine speed is compared with the actual engine speed, the feedback amount is determined by multiplying the deviation by a proportional coefficient, and the feedback amount is superimposed on the command injection amount. Idle feedback means so that the number approaches the target engine speed, and the idle feedback means learns larger than the pre-learning proportionality coefficient for before learning and the pre-learning proportionality coefficient for after learning. 3. The fuel injection according to claim 1, wherein a post-learning proportional coefficient is stored, the pre-learning proportional coefficient is used before learning, and the post-learning proportional coefficient is used after learning, and the feedback amount is determined using the post-learning proportional coefficient. Control system. When the learning is completed, the correction amount learning unit stores in the memory whether or not learning has been performed together with the learned correction amount, and performs learning when it is determined that learning has been performed based on the data indicating whether or not learning has been performed. Instead, the command injection amount determination means determines whether it is before learning or after learning based on the data indicating whether learning has been completed, and whether or not a re-learning condition set in advance is satisfied. Re-learning determination means for determining whether or not when the re-learning determination means has become a re-learning condition, the data indicating whether or not the learning has been completed is set to a state before learning. The fuel injection control system according to any one of claims 1 to 3, further comprising a learning advance return means for returning the correction amount learning means and the command injection amount determination means before learning.

Images